Use grep -oP to parse regexes on amazon linux machines

test/metric_value_benchmark/jmx_kafka_test.go

@@ -6,6 +6,7 @@
 package metric_value_benchmark
 
 import (
+	"fmt"
 	"log"
 	"time"
 
@@ -53,10 +54,16 @@ func (t *JMXKafkaTestRunner) SetupBeforeAgentRun() error {
 		return err
 	}
 
+	log.Println("get latest kafka version")
+	version, err := common.RunCommand("curl https://dlcdn.apache.org/kafka/ | grep -oP \"\\d\\.\\d\\.\\d\" | tail -1")
+	if err != nil {
+		return fmt.Errorf("error getting latest kafka version: %v", err)
+	}
+
 	log.Println("set up zookeeper and kafka")
 	startJMXCommands := []string{
 		"export KAFKA_VERSION=curl https://dlcdn.apache.org/kafka/ | grep -oE '\\d\\.\\d\\.\\d' | tail -1",
-		"curl https://dlcdn.apache.org/kafka/$KAFKA_VERSION/kafka_2.13-$KAFKA_VERSION.tgz -o kafka_2.13-latest.tgz",
+		fmt.Sprintf("curl https://dlcdn.apache.org/kafka/%s/kafka_2.13-%s.tgz -o kafka_2.13-latest.tgz", version, version),
 		"tar -xzf kafka_2.13-latest.tgz",
 		"echo 'export JMX_PORT=2000'|cat - kafka_2.13-latest/bin/kafka-server-start.sh > /tmp/kafka-server-start.sh && mv /tmp/kafka-server-start.sh kafka_2.13-latest/bin/kafka-server-start.sh",
 		"echo 'export JMX_PORT=2010'|cat - kafka_2.13-latest/bin/kafka-console-consumer.sh > /tmp/kafka-console-consumer.sh && mv /tmp/kafka-console-consumer.sh kafka_2.13-latest/bin/kafka-console-consumer.sh",